I got some output that had a “tool_calls” but the finish_reason is stop which isn’t great cause eariler I was checking if there was a tool_calls in my code to do different behavior based on the finish_reason.
I can’t tell if this is a bug cause there’s no natural stop point here cause there’s no content.
How do you reliably check if there’s a tool_calls in the api ?
I’m using python SDK btw.
{
"id": "chatcmpl-9a6ZoOIgaT9AjIQO1YU7UmMajiX39",
"choices": [
{
"finish_reason": "stop",
"index": 0,
"logprobs": "none",
"message": {
"content": "none",
"role": "assistant",
"function_call": "none",
"tool_calls": [
{
"id": "call_vwpeVemUh8Xk34gXncMaHXGt",
"function": {
"arguments": "{"location":"Sunriver, OR","query":"schools"}",
"name": "google_nearby_places"
},
"type": "function"
}
]
}
}
],
"created": 1718392116,
"model": "gpt-3.5-turbo-0125",
"object": "chat.completion",
"system_fingerprint": "none",
"usage": {
"completion_tokens": 21,
"prompt_tokens": 478,
"total_tokens": 499
}
}